from Western Szechueii. 375 



scarcely be regarded as distinct. When carefully compared 

 it appears that the male of C. dubius is somewhat greyer in 

 tlie brown of its upper parts, and slightly deeper in the red 

 of its underparts. 



I am inclined to think that C. thura is confined to the 

 Himalayas, and that the pair of birds which FAbbe David 

 obtained in Western China, and identified as Propasser thura, 

 ought to have been referred to C. dubius. 



-f- Carpodacus rubicilloides. 



A female of this species in the collection agrees with other 

 females collected by General Prjevalski in Kansu. The rump 

 is not suffused with either red or yellow ; most of the feathers 

 have a broad dark shaft-streak, which is conspicuous on the 

 underparts, including the flanks and under tail-coverts, and 

 on the upper parts except on the rump and upper tail-coverts, 

 where it is very obscure. This brings us down to the bot- 

 tom of page 390 in Dr. Sharpens key to the species in the 

 twelfth volume of the ' Catalogue of Birds in the British 

 Museum,^ w^here it completely breaks down. C. roseipectus 

 has probably been inserted by accident, as it belongs to the 

 smaller group on the next page, and brings the key to a 

 premature end, where it now stands. The characters which 

 follow are of no value : '' wing over 3'6 inches " will not include 

 any female of C. grandis in my collection, and does not 

 include the female of C. rhodochlamys described in the 

 Catalogue. I quite agree with Dr. Sharpe [op. cit. p. 407) 

 where he says that the female of C. rhodochlamys is not 

 distinguishable from that of C. grandis, and consequently I 

 disagree with his assertion {op. cit. p. 391) that the former 

 differs from the latter in having the rump streaked with 

 blackish brown. The nearly uniform rump is characteristic 

 of C. ruticilla and C. severtzovi, which differ from each other 

 in their under tail-coverts, the former species having very 

 broad and the latter very narrow dark centres to these 

 feathers. These two species are larger (wing4| to 4| inches); 

 C. rubicilloides occupies a medium position (wing 4| to 3| 

 inches) ; whilst C grandis and C. rhodochlamys are smaller 

 (wing '6\ to 3^J inches). 



